Large rise in carbon uptake by land plants


A proxy for the amount of carbon dioxide taken up by plants for photosynthesis has been used to estimate historical global uptake, revealing a large increase that might partly offset the rise in atmospheric CO2 levels. See Letter p.84

The rate of increase of atmospheric concentrations of carbon dioxide largely reflects the balance between anthropogenic emissions and natural uptake by the biosphere and the oceans1. Photosynthesis by land plants accounts for the greatest uptake of CO2, but this flux cannot be measured directly, and estimates2 of the changes in the flux during the twentieth and twenty-first centuries have varied enormously — between +5% and +52%. On page 84, Campbell et al.3 report that measurements of atmospheric levels of a gas called carbonyl sulfide (COS) can be used as a tracer of global changes in photosynthetic CO2 uptake. The authors thus provide a much-needed global observational constraint on historical increases in this uptake, which they argue must be at the high end of previous estimates.

Human activities release nearly 10 billion tonnes of carbon into the atmosphere each year4. Just less than half of this amount remains in the atmosphere, causing atmospheric CO2 concentrations to rise. The rest is removed by the land biosphere and the oceans4. Evidence suggests4,5 that the amount of CO2 taken up for photosynthesis by land plants — a quantity known as gross primary production (GPP) — has increased during the industrial period.

However, GPP can be estimated only indirectly, either from models or from ecosystem-scale measurements that capture the small difference between the opposing fluxes of photosynthetic uptake and emissions from respiration. GPP estimates based on models differ substantially from those based on data, and even estimates from different model simulations vary greatly2. Motivated by this serious problem, Campbell et al. introduce an approach to reducing the uncertainty, using records of long-term changes in the atmospheric concentrations of COS.

COS is a sulfur-containing analogue of CO2. It has a lifetime of 2–3 years, and mean atmospheric concentrations of about 500 parts per trillion (p.p.t.)3,6. Interest in this gas among carbon-cycle researchers grew in 2007 when a paper6 reported seasonal, cyclic changes in its atmospheric concentrations (decreases in summer and increases in winter), and that these variations are greatly diminished in the Southern Hemisphere. These changes are evidence of the 'breathing' of the land biosphere — they correlate with the seasonality of CO2-associated plant photosynthesis and respiration in the land biosphere of the Northern Hemisphere.

It is now generally accepted that the seasonal cycle of COS is intricately linked to that of CO2, but that there are key differences7 (Fig. 1). First, the COS cycle reflects COS production predominantly in the surface ocean, and consumption almost exclusively by land plants during — but not as part of — photosynthesis. Second, unlike CO2, COS is destroyed in leaves because it undergoes an irreversible hydrolysis reaction, so that COS fluxes reflect the one-way gross flux into leaves. GPP can therefore be calculated from COS measurements using a COS-to-CO2 'currency converter'8. However, such calculations come with caveats, because some aspects of COS dynamics (such as uptake and production of the gas by soil, and its consumption by plants at night) are still topics of research.

Figure 1: Global cycling of carbon dioxide and carbonyl sulfide (COS).

CO2 exchange between the atmosphere and either the oceans or the land biosphere reflects the balance between several opposing fluxes of CO2 uptake and release. This makes it difficult to use atmospheric CO2 measurements to quantify photosynthetic uptake by land plants. By contrast, COS is predominantly produced in the ocean surface from the decomposition of sulfur-containing organic matter, and is almost exclusively consumed by land plants in processes that accompany photosynthesis. Because the ratio of COS to CO2 uptake by land plants during photosynthesis is predictable, Campbell et al.3 were able to use records of atmospheric COS concentrations for the past 54,300 years to show that the global rate of CO2 flux into plants increased greatly during the twentieth century. CO2 and COS (or its precursors) are also produced in large quantities by human activities, and some CO2 and COS can be exchanged between the atmosphere and the soil.

To scrutinize the wide range of previous estimates of GPP changes, Campbell et al. used records of past changes in atmospheric COS levels based on measurements of air bubbles trapped in Antarctic ice and firn (snow deposited in past seasons, an intermediate stage between snow and glacial ice), and on atmospheric and satellite data. Their analysis shows that COS levels were stable for most of the 54,300-year record, remaining below 350 p.p.t., but then increased sharply during the industrial period (approximately the past 200 years). Concentrations peaked at about 550 p.p.t. in the late 1980s, then dropped, eventually stabilizing at around 500 p.p.t.

The authors next performed numerical simulations to see whether they could reproduce the observed atmospheric record, using published estimates of all the production and consumption terms that contribute to the atmospheric COS budget. Their analyses indicate that the simulations are most sensitive to the particular ocean and anthropogenic sources of COS used, and to plant consumption of COS. The authors argue that the ocean source would not be expected to have changed markedly during the industrial period, and so anthropogenic emissions and plant consumption must account for most of the observed COS changes in that period. They conclude that a large increase in plant-uptake flux between 1900 and 2013 is required in the simulations to obtain a good agreement with the observed atmospheric COS record for that period, irrespective of the anthropogenic source chosen for the simulations. This increase corresponds to GPP growth of 31 ± 5%, which, in turn, corresponds to some of the highest previously reported estimates of historical GPP growth.

The authors' study provides a badly needed observational constraint on the global growth rate of GPP and pioneers the use of COS in this context. However, the findings are mainly based on the correlation of a complex atmospheric COS record with simplified model simulations. It is therefore clearly not the final word on the topic. In particular, further efforts will be needed to work out the processes underlying the increase in GPP.

Enhancement of photosynthesis during the industrial period has been ascribed to the 'fertilizing' effects of rising atmospheric CO2 levels9 and of increased deposition of nitrogen-containing compounds from the atmosphere10 (associated with emissions from fossil-fuel combustion and intensification of agriculture). Global warming and land-use management11 are also thought to have had a role. Some of these factors might have been underestimated and will need to be reassessed. Given that changes in atmospheric CO2 levels are accurately measured and reflect the balance of all fluxes between Earth's surface and the atmosphere, any alterations to estimates of one of those fluxes (such as GPP) will require adjustments to others. Changes in the respiration of the biosphere and ocean processes, for example, might therefore need to be independently addressed.

Many scientists agree that we cannot rely only on CO2 measurements to unravel the CO2 cycle. By adding COS to our limited toolbox, Campbell et al. have taken a crucial step forward for those working in this field. The results will inform climate predictions, which currently are unable to determine whether the terrestrial biosphere and its photosynthetic capacity (GPP) will continue to take up more than one-quarter of anthropogenic CO2 emissions. Footnote 1


  1. 1.

    See all news & views


  1. 1

    Le Quéré, C. et al. Earth Syst. Sci. Data 7, 349–396 (2015).

  2. 2

    Friedlingstein, P. et al. J. Clim. 27, 511–526 (2014).

  3. 3

    Campbell, J. E. et al. Nature 544, 84–87 (2017).

  4. 4

    Raupach, M. R. Nature Clim. Change 1, 148–149 (2011).

  5. 5

    Ballantyne, A. P., Alden, C. B., Miller, J. B., Tans, P. P. & White, J. W. C. Nature 488, 70–72 (2012).

  6. 6

    Montzka, S. A. et al. J. Geophys. Res. 112, D09302 (2007).

  7. 7

    Launois, T., Belviso, S., Bopp, L., Fichot, C. G. & Peylin, P. Atmos. Chem. Phys. 15, 2295–2312 (2015).

  8. 8

    Asaf, D. et al. Nature Geosci. 6, 186–190 (2013).

  9. 9

    Schimel, D., Stephens, B. B. & Fisher, J. B. Proc. Natl Acad. Sci. USA 112, 436–441 (2015).

  10. 10

    Bala, G., Devaraju, N., Chaturvedi, R. K., Caldeira, K. & Nemani, R. Biogeosciences 10, 7147–7160 (2013).

  11. 11

    Arneth, A. et al. Nature Geosci. 10, 79–84 (2017).

Download references

Author information

Correspondence to Dan Yakir.

Related links

Related links

Related links in Nature Research

Biogeochemistry: As different as night and day

Biogeochemistry: Carbon dioxide and water use in forests

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Yakir, D. Large rise in carbon uptake by land plants. Nature 544, 39–40 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.